Wallet Risk Scoring on Base cohosted with Base, Hyperbolic, Virtuals and Pond. For more details and submit your models, please visit: https://cryptopond.xyz/modelFactory/detail/65 Assess wallet risk levels by analyzing transaction histories, wallet-to-wallet transfers, and other on-chain activities. Overview In the blockchain ecosystem, transactions occur in a decentralized, anonymous manner. While this feature provides freedom and security, it also enables fraudulent activities like money laundering, Sybil attacks, and other illegal operations. For decentralized finance (DeFi) organizations, exchanges, and regulators, identifying and assessing wallet risk has become a critical challenge. While blockchain provides transparency by publicly recording all transactions, interpreting these transactions presents significant challenges. Fraudulent activities often hide within millions of daily transactions, and wallets involved in illegal activities can undermine blockchain protocols' stability and reputation. This situation demands an automated, efficient tool to evaluate wallet risk levels based on transaction history and on-chain behavior. Currently, determining a blockchain wallet's risk level lacks quick, reliable methods. Traditional manual or rule-based approaches prove inadequate and can't scale to handle large transaction volumes. This limitation creates several risks: Delayed detection of illegal transactions Financial losses for DeFi organizations and exchanges due to high-risk wallet interactions Non-compliance with AML (Anti-Money Laundering) and KYC (Know Your Customer) regulations Inadequate risk management of blockchain wallets threatens both ecosystem security and legal compliance. A wallet risk scoring system would: Improve security: Reduce illegal and fraudulent blockchain transactions Support regulatory compliance: Ensure blockchain organizations meet AML and KYC requirements Build Trust: Strengthen user confidence in blockchain and DeFi services Optimize operations: Help organizations save time and costs when checking wallet legitimacy If you are unfamiliar with the basic concepts in Crypto such as tokens and wallets, please start with our blog post "Blockchain 101". Otherwise, let's dive in! Below is the competition process. The online phase of this competition runs for 7 weeks. During this period, you may submit your model results up to 3 times per day. Half of the records in your submission file will be used for initial evaluation, with preliminary scores displayed on the Leaderboard. At the end of week 7, you must submit your best model result before the submission portal closes. The remaining half of the records in your best submission will be used to determine the final score. The top 10 model developers will be notified and advance to the second round, where they'll present their work to a committee of AI leaders from the competition hosts. If notified, please respond within 48 hours or you lose your spot to the next-ranked developer. During week 8, these finalists will prepare their presentations and upload their models according to Pond's technical specifications. Presentations occur at the end of week 8, with final rankings determined by both evaluation scores and committee feedback. The top 3 developers will receive their prizes after their models are successfully verified and deployed to Pond Models platform. The prizes are: 1st place will be awarded $4000, 2nd place $2000, and 3rd place $1000. Additionally, revenue from the models’ future usage will be distributed back to their developers proportionally.